Back to Marketplace
Development toolsPlugin

dsh-eval-harness

BiBoyang/dsh-eval-harness

dsh-eval-harness is a tool for DeepSeek Harness. DSH plugin eval tool: YAML case-driven real agent regression eval + baseline comparison PASS/WARN/FAIL gates|Regression eval harness for DeepSeek Harness plugins

BiBoyang/dsh-eval-harness Screenshot preview
Stars
13
Forks
1
Issues
1
Update
7 days

REVIEW SCORECARD

Review Scorecard

7.0/10Average
License compliance6.0
Engineering quality6.2
Content completeness7.0
Repository scale7.0
Maintenance activity9.0

✓ No risk detected · Automated scan results are for reference only, not an official endorsement

AI DEEP REVIEW

Worth AI Review

7.8/10Recommended

A well-engineered regression eval harness for DSH plugins with real-agent tracing, baseline gating, and CI exit codes; w

Code quality8

README documents a zero-dependency YAML subset parser, strict schemaVersion validation with legacy fallback, per-attempt result history, and unbiased pass@k/pass^k estimators rather than the biased (c/n)^k, indicating deliberate engineering rigor.

Security7

The tool spawns local dsh subprocesses with isolated session roots and workspaces, but the LLM judge sends outputs to an OpenAI-compatible endpoint configured via EVAL_JUDGE_API_KEY/DEEPSEEK_API_KEY environment variables, so no plaintext secrets or download-and-execute patterns are visible.

Utility8

It addresses a concrete, common pain point for plugin and skill authors by turning YAML cases into real-agent regression runs with PASS/WARN/FAIL gating and CI exit codes, plus a judge-calibration tool reporting TPR/TNR.

Maintenance8

Last updated 2026-08-26 with an active repository, a documented deprecation of the stale npm channel in favor of GitHub distribution, and schema versioning that anticipates future changes.

Documentation8

The README covers installation with a mount-verification command, full tool and skill tables, an annotated YAML case format, parameter tables with defaults, and explicit parsing limitations, though the excerpt is truncated mid-section on eval_judge_validate.

Generated by AI after reading the project README — a decision aid, not an endorsement. Neutral scores are given when information is limited.

PROJECT READMEREADMEExpandCollapse

Loading GitHub README...

INSTALL REFERENCE

Install reference

Structure check pending
dsh plugin --profile headless add github:BiBoyang/dsh-eval-harness

This command was extracted from an explicit README install reference and is display-only; the DSH host will not execute it automatically.

PROJECT TOPICS

Project topics

VALIDATION LADDER

Validation progress

Structure check pending
01Store discoveryPassed
02ClassificationPassed
03Structure checkPending
04Sandbox validationPending

FAQ

Frequently Asked Questions

What is dsh-eval-harness used for?

DSH plugin eval tool: YAML case-driven real agent regression eval + baseline comparison PASS/WARN/FAIL gates|Regression eval harness for DeepSeek Harness plugins

How popular is dsh-eval-harness?

dsh-eval-harness has 13 stars and 1 forks on GitHub, last updated 2026-09-25.

Is dsh-eval-harness compatible with DeepSeek Harness?

dsh-eval-harness is listed in the DSH Universe directory as a tool for DeepSeek Harness. This plugin is listed in the DSH Universe directory and covered by its validation pipeline.

CLASSIFICATION EVIDENCE

Classification evidence

Project typePlugin
Feature categoryDevelopment tools
Rule confidenceMedium

The system checks GitHub Topics against the site taxonomy and keyword rules. Current matches:regression-testing。